[Gdal-dev] new RFC 13: Improved Feature Insertion/Update/Delete Performance in Batch Mode

[Frank Warmerdam]

Daniel Morissette wrote:
> I was not aware that OGR had this transaction mechanism, but I agree 
> with Frank that it would be preferable to it if possible instead of 
> adding another mechanism to do the same thing.

Daniel,

I suspect that you don't know about the transaction mechanism because it
is only lightly supported and advertised.  In fact, I suspect that the approach
as it currently stands is a bit poorly thought out.  For instance, transactions
are datasource wide, rather than per-layer which causes problems in some
circumstances.  I believe there are substantial patches for the postgres
driver to address flaws in the transaction implementation there that I'm afraid
to apply for fear of what else it might break.

And very few providers (even RDBMS providers) bother to implement the
transaction semantics.

Best regards,
-- 
---------------------------------------+--------------------------------------
I set the clouds in motion - turn up   | Frank Warmerdam, warmerdam at pobox.com
light and sound - activate the windows | http://pobox.com/~warmerdam
and watch the world go round - Rush    | President OSGeo, http://osgeo.org